Water & Wastewater Treatment

34 | november 2014 | WWT | www.wwtonline.co.uk In the know Research round up: pipes and pipelines and the complete facility is tempera- ture controlled. Each loop includes a number of uniquely designed ports and coupons through which represent- ative samples can be collected from the inner surface. This facility let the UoS researchers perform a number of different experi- ments and analyses in an interdiscipli- nary collaboration between engineers, microbiologists and computer scien- tists. Early experiments involved recre- ating WDN accumulation and flushing phases. They used a combination of turbidity and WQ indicators, alongside inspections of the inner surface, in order to determine how the tenacity of the biofilm is affected by variations in the hydraulic regime, and extent to which interventions dislodge biofilm build-up from the internal surface. They proved that the "community structure" of the biofilm – the specific species and families of microbes to be found within it – is an important factor, as it influences the evolution of the biofilm and the likelihood of problematic WQ events. Hundreds of biofilm samples were subjected to ge- netic sequencing in order to map the shi†ing catalogue of microbes which inhabit them. However, the whole point of Pipedreams is to take laboratory knowledge out into the field, and ap- ply it to real problems. The laboratory experiments were complemented by fieldwork done in partnership with nu- merous water companies. These stud- ies of real systems in action provided the foundational data-sets with which the team's computer scientists have built and refined a variety of models and scenarios for forecasting WQ risk. Findings Pipedreams has generated masses of results, as indicated by sixteen journal papers (so far), and a similar number of conference papers – far too much to cover here, and much of it very specialised. The really exciting stuff, however, emerges when the results from different disciplines combine to produce new knowledge that's action- able by network managers in the field – and this is where the interdis- ciplinary Pipedreams approach, and through the related PODDS project- sequence, has paid dividends. For example, it's now understood that flush interventions never remove all the material from the inner surface, and that the hydraulic factor which most closely governs biofilms and associated WQ risk is the daily peak flow in the system, as opposed to total daily flow or the nocturnal stagnation. Indeed, flushing appears to selectively mobilise the biofilm community struc- ture, with some species better able to weather the change in conditions. Furthermore, biofilm build-up can be retarded by establishing a hydraulic regime of low temperatures and varied flow; however, such a regime also tends to produce more sturdy biofilms which resist flushing. The hydraulic regime is more influ- ential on the physical tenacity of the biofilm than on its community struc- ture, while the latter varies in relation to the pipe material; due to pitting and corrosion, iron pipes offer appealing microenvironments for some particu- larly stubborn microbial species with a well-established connection to iron ac- cumulation and discolouration events. These insights into the relationship between biofilm layer behaviour and hydraulic conditions, in particular shear stress, resulted in the develop- ment of a model which extrapolates WQ risk into "scenarios", which are explored mathematically to develop robust strategies for the safe and flexible management of legacy WDNs, especially in combination with the deployment of real-time sensors in live networks. Also under development are "self-organising maps", a sort of neu- ral network so†ware which clusters data together in order to determine how different factors in a system are interrelated; these are, for example, being applied to the question of biofilm growth rates, and how they relate to factors such as pipe material or water-source type. Applications & Next steps Pipedreams has advanced understand- ing in multiple disciplines, but it's the combination of these new knowledges and tools that promises to change the way the water sector plans for the future. Pipedreams provides sim- ple, actionable answers for network managers seeking to improve service quality while keeping costs down. For example, the discolouration risk model offers a flexible and strategic toolkit for asset management which would allow interventions – from flushing to out- right asset replacement – to be targeted for maximum effect at minimum cost. But Pipedreams also opens up the possibility of live WQ monitoring and real-time system management from the control room. As the cost of monitoring instruments falls and their quality in- creases, the Pipedreams models show the potential for robust detection and diagnosis of WQ events in live, critical networks. By reducing the need for on-site inspections, supply stoppages and other costly interventions, such an investment would soon pay itself off with significant reductions in both operational expenditure and customer complaints. The Pipedreams research project was sponsored by the engineering & Physical Sciences research Council under its Challenging engineering scheme. Industry partners included Anglian Water, Dwr Cymru Welsh Water, northumbrian Water, Scottish Water, Severn Trent Water, United Utilities, Wessex Water and Yorkshire Water. Find out more about Pipedreams and PoDDS at www.shef. ac.uk/pipedreams and www.podds.co.uk Biofilms observed through a microscope a er 28 days of growth in a pipe system

• Cover